JPH0457619A - Discharge gap control device - Google Patents

Abstract

PURPOSE:To maintain a stable discharge gap for a long time by providing a central processing unit for inferring and operating optimum output from a deviation and the time change rate thereof using the membership function of a memory device, and executing fuzzy inference for outputting an operation amount via the synthesis of each inference result. CONSTITUTION:The control of a discharge generation gap is made in such way that gap voltage or the like proportional to a discharge gap 3 is detected with a detector 5. Furthermore, a deviation resulting from the comparison of the detected voltage or the like with the reference voltage of a setting element 7 in a comparison operation circuit 6, and the time change rate E thereof proportional to a feed rate are computed, and inputted to a fuzzy control section 8 as a fuzzy amount. An operation amount inferred from the amount through a membership function is used to control a discharge gap or the drive control of a machining feed device 4. According to the aforesaid construction, the occurrence of overshoot or undershoot can be reduced, and gap control can be stably done in discharge machining or the like giving a very complicated fluctuation or behavior. In addition, it is possible to control the gap very strongly and stably against disturbance due to a chip or the like.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention generates electrical discharge between opposing electrodes in electrical discharge machining, wire electrical discharge machining, micro welding, racer processing, xenon discharge, etc., and utilizes the electrical discharge. This invention relates to a discharge gap control device used in equipment that performs machining.

[Problems with conventional technology]

For example, in an electric discharge machining apparatus, control is performed such that the electrode is servo-followed to the workpiece so that the discharge gap is always constant.

Conventionally known electrode servo machining feed control devices detect signals such as voltage and current that are proportional to the discharge gap, and directly control a machining feed device such as a servo motor based on the difference between this signal and a preset reference value. The discharge gap is controlled to a predetermined optimum value.

[Problems to be solved by invention]

However, machining debris, gas, etc. generated during machining are present in the discharge gap, and the gap detection signal fluctuates depending on the behavior of this machining debris, and the servo motor drive control is performed in response to this fluctuation. Therefore, with this method, it is not possible to maintain a predetermined stable gap for a long period of time.

If the gap narrows below a certain value due to overfeeding of the servo, arc discharge will easily occur and stable machining will not be possible, so the reference voltage for servo control should generally be set high and always work within the safe range. However, if the reference voltage is set high, the discharge gap will widen, resulting in a decrease in the rate of occurrence of discharge, and a decrease in machining efficiency, that is, machining feed rate.
This may lead to the problem of longer processing time.

An object of the present invention is to improve the gap control device having such problems and to provide a new control device that can maintain a stable gap over a long period of time.

[Means to solve the problem]

The object of the present invention is to provide a detection device for detecting the state of the discharge gap, and to compare the detection signal of the detection device with a set reference value to detect the deviation E.
and its temporal change rate ΔE, as well as the required amount of fuzzy rules, membership functions, calculation programs, etc. set for the above deviation E and its temporal change rate ΔE, and the output U variable. a storage device written with
Input E and ΔE using the membership function of the storage device above.
a central processing unit that executes fuzzy inference that calculates the optimal output U from each inference, synthesizes the results of each inference, and outputs a manipulated variable, and supplies the output of the central processing unit to the machining feed device for controlling the discharge gap. This is achieved by means of a discharge gap control device comprising a fuzzy control circuit comprising an interface.

[Effect]

The present invention controls the discharge generation gap by detecting the gap voltage, etc., which is proportional to the discharge gap, and comparing this with a reference voltage.
and its temporal rate of change ΔE, which is proportional to the feed speed, are input as ambiguous quantities, and the discharge gap is controlled using the manipulated variable obtained by fuzzy inference using the membership function. Because of this, compared to conventional feedback control, there are fewer overshoots and undershoots, and rapid convergence control to the set value is possible, making gap control such as electrical discharge machining, which has extremely complex changes and behaviors, stable. It is extremely resistant to disturbances caused by machining debris, etc., and can perform stable control.

Therefore, according to the present invention, electric discharge machining, micro welding welding, racer machining, etc. can be performed safely and reliably with high precision and high speed.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an electric discharge machining apparatus equipped with a discharge gap control device according to the present invention, FIG. 2 is a graph showing an example of membership functions set in memory, and FIG. 3 is an explanatory diagram of fuzzy inference. Figure 4 is an explanatory diagram of the combined output of the inference results, Figure 5 is a diagram in which membership functions are written as discrete values, and Figure 6 is a configuration diagram showing an example of modification of the Fushida electrical discharge machining equipment to Figure 1. be.

In FIG. 1, l is an electrode, 2 is a workpiece, 3 is a discharge gap formed between the electrode 1 and the workpiece 2, 4 is a machining feed device such as a servo motor, and 5 is a workpiece. A detector 6 for detecting the average voltage of the discharge gap compares the detection signal voltage with a preset reference voltage, and calculates the deviation value E and its temporal change rate ΔE, that is, the deviation during one sampling period. Comparison calculation circuit for calculating the amount of change, 7 is a reference voltage setting device, 8
9 is a fuzzy control section, and 9 is an amplification waveform shaping circuit.

Pulses are supplied between the electrode 1 and the workpiece from a known machining power source (not shown), and the machining is fed toward the electrode 1 or the workpiece 2, thereby generating a pulse discharge between the two. , electrical discharge machining of the workpiece 2 is performed by the electrode 1.

The detection voltage is the average voltage across the gap and is proportional to the gap.

A reference voltage value corresponding to the optimum discharge gap is set in the reference voltage setter 7, and the comparison calculation circuit 6 calculates the deviation E of the detection signal and outputs it as an ambiguous quantity for fuzzy inference.

Similarly, the comparison calculation circuit 6 determines the rate of change over time ΔE of the deviation E.

This rate of change ΔE corresponds to the feed speed, but this rate of change ΔE is also output as an ambiguous quantity.

The deviation E and rate of change ΔE from the comparison calculation circuit 6 are input to the fuzzy control section 8.

The fuzzy control unit 8 includes a CPU (central processing unit) 61,
It consists of a memory 62, interfaces 63, 64, 65, etc. As the comparison calculation circuit 6, for example, an 8-bit one-chip microcomputer or the like can be used.

The memory contains data necessary for fuzzy control programs and operations such as fuzzy rules, membership functions, and inference operations.

The following rules are set as fuzzy rules for discharge gap control.

A is a gap or ten people, B is underground, C is a bank, D is a predetermined value, E
is set to one small, F is set to one medium, and G is set to one large, and the change in the operation amount for each is A → Large and fast in the direction of narrowing the gap B - Medium speed C to medium in the direction of narrowing the gap → Small and slow speed in the direction of narrowing the gap D → Stop as it is E + Small and slow speed in the direction of widening the gap F → Medium speed G in the direction of widening the gap → Large and fast speed in the direction of widening the gap.

As the membership function, a triangular type having continuous values as shown in FIG. 2 was used, and the function was provided in the range of ±30V of the reference voltage.

That is, as shown in the figure, the triangular center A is set to +30V, B is set to -30V, and the center D is set to a function of Ov, and seven triangles slightly overlapping each other were adopted.

The following rule is used as a control rule for finding the optimal output value U using input values E and ΔE.

R+; xf E Is A andΔE is
A then U is thicker in the direction of narrowing the gap (faster R49; if E is G and ΔE is G
then U is thicker in the direction of widening the gap (quickly, here A indicates a state where the gap is very wide, G indicates a state where the gap is very narrow, and E is larger than ΔE)
If is very large, the rule is to operate U rapidly and greatly in the direction of narrowing the gap.

Since each of the two input values of E1ΔE has 7 fuzzy sets, 7×7 rules can be created for the control rule.

The discharge gap voltage is detected by the detector 5, and the detected signal is compared with a set reference value by the comparison calculation circuit 6 to calculate the deviation E and its rate of change ΔE.

These E and ΔE are input as ambiguous quantities to the fuzzy control unit 8 for fuzzy inference.

The information tK input through the interfaces 83 and 84 is stored in an accumulator inside the CPU 81, and fuzzy inference calculation processing is performed based on the contents of this accumulator.

now, E=E.

ΔE=ΔE.

Then, inference can be started by finding the grade.

The grade of the antecedent part is the smaller of E and ΔE because there is an and in the rule.

In Figure 3, rule R1 has a smaller grade of V.
, then by multiplying this Vl by 1A of the fuzzy set of U specified by the consequent part, the fuzzy set V+xA is obtained.

Also, in rule R2, if the grade is V2, a fuzzy set V2 XA is obtained.

In this case, it is possible to perform safe control by adopting the smaller grade and taking the safety factor.In this way, similar calculations are made for all fuzzy sets to which E and ΔE belong, and each rule The obtained inference results are combined to obtain the manipulated variable U.

In FIG. 4, considering all cases, the center of gravity of the union created based on the inference results of each rule is taken, and the operation amount U is determined by the center of gravity method.

The center of gravity U is the inference result.

The manipulated variable U obtained through fuzzy reasoning in this way is outputted through the interface 65, and the output signal is converted into a predetermined signal by the amplified waveform shaping circuit 9, which is added to the servo control device 4 to control the gap 3. .

That is, arithmetic processing of fuzzy inference is performed in response to the input, and an operation signal for fuzzy control is generated. Based on this command signal, the drive control device for gap control applies a pulsed or analog control signal to the processing feeder 4 and starts processing. The feed is controlled so that the discharge gap rapidly converges to the optimum value.

The processing feed device 4 is a pulse motor, a DC motor, a hydraulic cylinder, or the like.

In this way, changes in the state of the discharge gap are constantly detected and information is obtained, fuzzy inference is made to the manipulated variable U corresponding to the ambiguous input value, and the weighted manipulated variable is output to perform gap control. Very rapid and responsive control of the discharge gap is achieved.

In addition, the above explained the servo control of the discharge gap in the case of electric discharge machining, but in the case of electric discharge machining, when machining debris accumulates in the discharge gap, the electrode is moved reciprocally up and down, and the pumping action is used to control the inside of the discharge gap. A so-called electrode reciprocating motion is generally performed to perform the cleaning operation, but this reciprocating control can be performed in combination with the servo control using fuzzy reasoning while detecting the discharge state of the discharge gap.

Further, the machining pulses supplied from the machining power source and the machining fluid supplied from the machining fluid supply device can also be controlled in the same manner as the discharge gap control described above.

When laser processing is involved, the position control of the condenser lens, the control of the oscillator, etc. are similarly performed in parallel.

Signal detection in the discharge gap is not limited to voltage, but may also detect discharge current, gap impedance, sound waves generated with discharge, electromagnetic waves, etc., or a combination of these parameters may be used.

Note that there are membership functions with continuous values, such as bell-shaped and exponential types, which can be modified to have a higher probability by tuning the signal.

Furthermore, it is also effective to use the weighted average method without determining the center of gravity when combining the inference results to determine the manipulated variable U.

Figure 5 shows a table in which the membership functions are written as discrete numerical data. , -Medium, -Small, 0, Bank, Dochu, and 10 people are defined, and these are made into a vertical and horizontal distribution table, each with 1 on the O axis.
Three constraint rules were set, and the blank values in other areas were set in sequence so that they converged toward the center.

In this way, about 49 points of discrete numerical data are written in the ROM in the form of a table.

For controlling the discharge gap, more stable gap control can be achieved by inferring the optimal operation amount using the memorized discrete data and controlling the machining feed device based on the output of the selected operation amount.

The detected signal is compared with a reference value to obtain the deviation value E and its rate of change over time ΔE, and these are input as ambiguous quantities to the fuzzy controller for fuzzy inference.

In the table shown in FIG. 5, the deviation value E and its temporal change rate ΔE are selected from data existing at the intersection of the corresponding columns.

This selected discrete value is used as the fuzzy inference grade for the antecedent part, and by multiplying it by a certain variable specified in the consequent part, the inference result is obtained.

In this way, inference using discrete data is easy, and by inputting a large amount of numerical data, more precise and more probable inference can be made, and the operation signal obtained in this way can be By outputting through the interface and controlling the processing feed device, even more stable gap control can be achieved.

The time interval for fuzzy inference is usually about 0.1 ms to several ms when controlling the gap in electrical discharge machining, and by performing fuzzy inference at intervals of this order and repeating the discharge gap control cycle, the fuzzy inference can be stabilized over a long period of time. Gap control and machining can be continued.

Next, an example of fuzzy control applied to wire-cut electrical discharge machining will be described.

A commercially available fuzzy controller was used, and control was performed at a sampling frequency (calculation interval) of 0.3 m5ec.

As the membership function, a function of ±30V was provided for the difference from the reference voltage. A 0.3 grade BS wire was used as a wire electrode for processing, and a steel material with a thickness of 0.5 mm was processed. Distilled water with a specific resistance of 5×10 5 Ω was used as the processing fluid. The machining shape was a cross-shaped cutting die, and the machine of the present invention controlled the feed using fuzzy reasoning, and the machining speed was 140 mm2.
When machining was performed using the / image, extremely stable machining was achieved without any disconnection at the - times.

For comparison, other than conventional feedback control,
When processed under the same conditions as in the above example, it was confirmed that an average number of wire breaks occurred during the entire processing time.

Figure 6 shows that the detection of the state of the discharge gap is input as an ambiguous quantity for fuzzy inference at the beginning of machining and in the middle of machining, and the discharge gap is controlled by fuzzy control. This is an embodiment in which the process can be carried out smoothly.

In FIG. 6, the same reference numerals as in FIG. 1 indicate the same parts.

31.32 is applied to the side wall of the machining tank during electrical discharge machining,
A discharge state detector 7 is placed relatively far from the discharge gap 3, and 33, 34 is a discharge state detector directly attached to the electrode l.

These detectors include magnetic sensing coils or MR
using elements. This is due to the fact that in electrical discharge machining, the state of magnetic field generation due to electrical discharge changes during normal normal machining and when the gap narrows and an arc discharge occurs or approaches it. Detector 33 placed in
．． The detector 31.3 detects the constantly changing gap state on the order of μs by the detector 31.3 located away from the gap.
2. Detect states that change slowly as the machining progresses in ms units, and input these detection signals to the comparison calculation circuit 6 to detect states that change at high speed moment by moment and states that change as the machining progresses. Compute and output ambiguous quantities.

The comparison calculation circuit 6 has terminals A, B, and C1 for setting reference values.
. . . Various optimum reference values predetermined in accordance with the progress of machining, that is, changes in machining depth and machining area, etc., are input in advance, and the detection signals are compared and processed.

The ambiguous quantity obtained in this manner is input to the fuzzy control section 8 to perform fuzzy inference, and generate a fuzzy control signal having a manipulated variable based on the fuzzy inference.

It was confirmed that by controlling the discharge gap by processing at calculation intervals of about 0.2 ms, stable electric discharge machining could be performed without generating arc discharge from the beginning to the end of machining.

In addition to the magnetic sensing coil, the above-mentioned detection device may be of any type as long as it can detect the gap state that changes from moment to moment and the gap state that changes gradually as the machining progresses. be able to.

Further, in the above explanation, a trigonometric function is used as the membership function, but the membership function is not limited to this, and bell-type, exponential-type, etc. can be used as appropriate.

Further, the fuzzy rules are not limited to a set of seven, and the best rules can be set by simulation. Furthermore, in addition to the method of taking the center of gravity of the union, a weighted average method may be used to synthesize the inference results.

〔Effect of the invention〕

As described above, the present invention performs fuzzy control to control the discharge gap, that is, detects a signal corresponding to the size of the discharge gap, and compares the detected signal with a reference value to obtain a deviation value. E and its temporal rate of change (velocity) are input as ambiguous quantities, fuzzy inference is performed using membership functions, and the machining feed device is controlled using the synthesized and output manipulated variables to calculate the discharge gap. control, and furthermore,
By setting the inference interval to about ms, the discharge gap can rapidly converge to the set value, and compared to conventional feedback control, overshoot, undershoot, and
There is little hunting and the discharge gap can be optimally controlled with high reliability.

In addition, it is stable against disturbances such as machining debris and arcs that occur in the discharge gap, and the gap can be kept constant with little fluctuation. Therefore, electric discharge machining, which processes while repeating pulsed discharge within the discharge gap, is stable. If the present invention is applied to wire-cut electrical discharge machining, welding by micro welding, racer machining, etc., these processes can be performed stably.
It has the effect of high-speed, efficient, high-precision precision machining.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an electric discharge machining apparatus equipped with a discharge gap control device according to the present invention, FIG. 2 is a graph showing an example of membership functions set in memory, and FIG. 3 is an explanatory diagram of fuzzy inference. Fig. 4 is an explanatory diagram of the combined output of the inference results, Fig. 5 is a diagram in which membership functions are written as discrete values, and Fig. 6 is a configuration diagram showing an example of modification of the electric discharge machining apparatus shown in Fig. 1. . 1... Electrode 2... Workpiece 3... Discharge gap 4... Machining feeder 5... Detector 6... Comparison calculation circuit 7...Reference setter 8...Fuzzy control unit 9...Signal amplifier 11■ ○ +30 ios-

Claims (2)

[Claims] (1) The workpiece (2) and the electrode (
1) facing each other, actuating a machining feed device (4) to apply machining feed to at least one of the two facing each other, and generating a discharge between the two in a discharge gap control device. , a detection device (5) that detects the state of the discharge gap, and a comparison calculation circuit (6) that compares the detection signal of the detection device (5) with a set reference value to calculate the deviation E and its temporal change rate ΔE. ,
In addition, a storage device (8
2), a central processing unit (81 ) and an interface (83, 84) for supplying the output of the central processing unit (81) to the machining feed device (4) for controlling the discharge gap. . (2) The detection device for detecting the state of the discharge gap comprises a detector that detects the constantly changing state of the discharge gap and a detector that detects the state of the gap that gradually changes as machining progresses. The discharge gap control device according to scope 1.